St Augustine and All That: Remarks on the beginning of Philosophical Investigations

One of the many striking things about the beginning of the Investigations is that it is not absolutely clear where the work begins. Does the Preface form an integral part of the book? If not, what becomes of the motto, which after all was added to the first page of the typescript of the last version of the Investigations? As we know, the Nestroy motto was added in 1947 or later to replace an earlier motto taken from Hertz’s Principles of Mechanics.^[1] So one may wonder whether its inclusion was fully authorised by Wittgenstein – just as one may wonder whether the Preface, which was written for an intermediate version of the work, was really meant to be printed in the form in which it has come down to us.

But even if one leaves all these perplexities on one side and takes motto and Preface at face value, one may wish to argue that, in view of the fact that these items form parts of the preliminary or introductory material, they are not to be counted as constituting the beginning proper of the book. In the present context, I don’t want to assess the value of this argument, nor do I want to express an opinion on any of the questions mentioned so far. I do, however, want to suggest that it can be helpful to consider the book as beginning with PI, 1, and hence with the well-known quotation from Augustine’s Confessions, and Wittgenstein’s translation of this quotation.

This passage from the Confessions is alluded to in manuscripts written as early as 1931; and in the Brown Book (which was dictated in the academic year 1934 – 35) Wittgenstein takes the decisive step of opening the work with an explicit reference to this passage. There, he writes: ‏”‎Augustine, in describing his learning of language, says that he was taught to speak by learning the names of things.”^[2] This remark is followed by a comment similar to one we know from the beginning of the Investigations. In the earlier work, Wittgenstein says that “It is clear that whoever says this has in mind the way in which a child learns such words as ‘man’, ‘sugar’, ‘table’, etc. He does not primarily think of such words as ‘today’, ‘not’, ‘but’, ‘perhaps’.”

More than a year later, towards the beginning of the period of his long Norwegian retreat, Wittgenstein worked on a German version of the Brown Book, which he entitled “Philosophische Untersuchungen” – Philosophical Investigations. The beginning is similar to the one readers know from the English Brown Book, only that now a few lines from Augustine’s text are quoted in the original Latin. As is well known, after having completed two thirds or so of the Brown-Book revision Wittgenstein abandoned this work and proceeded to assemble a new work, also entitled Philosophische Untersuchungen, which may be regarded as the first draft of what we know as the first 188 remarks of the published Investigations we are familiar with. This manuscript was revised and typed in 1937, and it is this version which looks recognisably similar to the one we know from the printed work. Wittgenstein’s translation of Augustine’s words, however, was only added to the so-called intermediate version of 1944 – 45.^[3]

Here I should like to point out a couple of small differences between early versions of the Investigations, on the one hand, and the final version and hence the published book, on the other. The first difference may appear trifling, but in the light of later commentary it seems important to bear it in mind. The difference is this: that the second paragraph of remark 1 of PI consists of what used to be two paragraphs in earlier versions of the text. That is to say, in these earlier versions Wittgenstein made it clear that an extra step needs to be taken if one wishes to get from the picture yielded by the quotation taken from the Confessions to the idea whose roots can be found in that picture. In other words, some – and perhaps a good deal of – extra work is required to obtain the idea from the picture. The relevant paragraphs run as follows:

These words, it seems to me, give us a particular picture of the essence of human language. It is this: the words in language name objects – sentences are combinations of such names.

In this picture of language we find the roots of the following idea: Every word has a meaning. This meaning is correlated with the word. It is the object for which the word stands.^[4]

So it is clear that a line should be drawn between the picture suggested by Augustine and the tripartite idea derived from this picture. And once this line is drawn, one may find that the picture is too unspecific and vague to yield a claim that could be seen as the direct source of something that might deserve to be called a “philosophical” or “theoretical” view of any kind.

So, if (as Wittgenstein writes in a fairly early manuscript) his point in quoting Augustine was that of mentioning the views of an unprejudiced and clear-thinking man who flourished in a much earlier period than ours and was alien to our own notions and preconceptions,^[5] we are probably expected to respond by saying something along the following lines: ‏”‎If this is a view held by an unprejudiced observer of the world surrounding him, then we shall need a good deal of reflection and theoretical work to arrive at the rudiments of something attracting and permitting philosophical discussion.”

And if this response is on the right track, we may wonder whether the translation rendering das Wesen der menschlichen Sprache as ‏”‎the essence of human language” is not perhaps too specific in that it suggests that the picture conveyed by Augustine’s words is an essentialist one. This may be so. At any rate, the ‏”‎rudiments of something attracting and permitting philosophical discussion” are only presented to us by the idea whose roots we may discover in the Augustinian picture proper. But if this is right, we should be aware of what I should not hesitate to call a fact, viz. the fact that the Augustinian picture properly so-called could give rise to a number of competing views on the nature of human language. And surely among these views there might be some that would not inevitably call forth criticism from Wittgenstein or another philosopher thinking along similar lines.

Accordingly, paying attention to this small difference between earlier versions and the final version of the text of the Investigations may help us to rethink, and perhaps modify, our attitude towards the Augustinian picture itself. By the same token, it may contribute to seeing the views criticised by Wittgenstein himself, or by his commentators on his behalf, as views we are more or less naturally inclined to hold. This, of course, does not render these views immune to criticism, but it may help us to see them as something deeply rooted in our own nature. This, by the way, is a way of looking at things endorsed by Wittgenstein himself when, in his remarks on Frazer, he insists on noting that magic and ritual actions can be understood to reflect our most basic instincts and that magic, in its turn, is ‏”‎always based on the idea of symbolism and language”.^[6]

Now I should like to point out the second small difference between earlier and later versions of remark 1 of the Investigations. This difference concerns what is the concluding paragraph of this remark in the book as we know it. In earlier versions, however, this concluding paragraph did not form part of the first remark at all: in all earlier versions up to the intermediate version of 1944 – 45 this paragraph formed a remark of its own. It was only in the intermediate version that Wittgenstein changed the numbering and turned what used to be remark 2 into the last paragraph of remark 1. So, by helping myself to a piece of terminology Wittgenstein only introduces in remark 7, I might describe the situation by saying that one of the two parallel language-games forming remarks 2 and 3 is deprived of its status and included in the first remark, where it figures as the last of four paragraphs (not counting the translation of the quotation from Augustine).

At first glance, at any rate, this modification by inclusion will seem like an odd decision on Wittgenstein’s part. After all, these two language-games – the shopkeeper’s language-game and that of the builders – are clearly of central importance. So why shouldn’t they be allowed to continue to stand on the same level, as it were? The architecture behind Wittgenstein’s arrangement and numbering of his remarks seems almost to demand the earlier ordering. I think that here we are faced with a real puzzle; and it is only towards the end of my paper that I shall venture a tentative answer to this problem. Now I shall quote the language-game of the shopkeeper and discuss a few questions raised by it:

Now think of the following use of language: I send someone shopping. I give him a slip of paper marked ‏”‎five red apples”. He takes the slip to the shopkeeper, who opens the drawer marked ‏”‎apples”; then he looks up the word ‏”‎red” in a chart and finds a colour sample [Farbmuster – BFF: färbiges Täfelchen = colour chart] next to it; then he says the series of elementary number-words – I assume that he knows them by heart – up to the word “five”, and for each number-word he takes an apple of the same colour as the sample [BFF: Täfelchen = chart] out of the drawer. – It is in this and similar ways that one operates with words. – “But how does he know where and how he is to look up the word ‘red’ and what he is to do with the word ’five’?” – Well, I assume that he acts as I have described. Explanations come to an end somewhere. – But what is the meaning of the word “five”? – No such thing was in question here, only how the word ‏”‎five” is used.^[7]

As I said, the expression ‏”‎language-game” is only introduced in remark 7, but there can be no doubt that its use to classify the shopkeeper scenario is legitimate. After all, in later passages of the book Wittgenstein himself explicitly speaks of (2) and (8) as language-games. Accordingly, there seems to be no reason to refrain from using the word with reference to the shopkeeper example.

What may be confusing is the circumstance that in early parts of the book Wittgenstein calls (2) and (8) ‏”‎languages”.^[8] It almost seems that in a few cases he uses the words ‏”‎language” and ‏”‎language-game” interchangeably. It is important to note and remember this, as Wittgenstein’s repeated invitation to think of his examples as complete languages is very hard to make sense of if the term ‏”‎language” is understood as referring to things of the same type as German or English, whereas the idea of completeness can get some grip if the word ‏”‎language” is understood as referring to language-games. That is to say, if you take a language-game to be something like a model of a delimited use of linguistic expressions, and hence as something that may be characterised by a kind of scenario or choreography, the question of completeness can begin to make sense. Analogous considerations apply to the idea of a fragment of language.

But in proposing this kind of characterisation I may seem to be rushing ahead. Many ways of reading Wittgenstein’s notion of a language-game have been put forward and defended in the literature, and there is no obvious agreement on a particular interpretation. So we shall need to add something by way of justifying the characterisation just given. In the present context, I shall confine myself to mentioning just a couple of points.

My first point is connected with a serious problem of translation. It has often been mentioned that the German word ‏”‎Spiel” can be difficult or even impossible to translate into English. It is obviously much wider than standard uses of the English word ‏”‎game”, covering sports as well as board or card games but not ring-a-ring o’ roses or other relatively aimless activities such as throwing a ball against a wall and trying to catch it, which is one of the Spiele explicitly mentioned by Wittgenstein (PI, 66).

I remember having discussions about the question whether it would not be better to translate the term Sprachspiel as ‏”‎language-play” rather than the accustomed ‏”‎language-game” in order to capture at least part of the extra meaning expressed by the German word ‏”‎Spiel”. It was in this context that people would bring up the inevitable reference to Friedrich Schiller’s letters On the Aesthetic Education of Man and his passionate claim that ‏”‎Man plays only when he is in the full sense of the word a man, and he is only wholly Man when he is playing”.^[9] I am not sure whether Schiller’s emphatic statement is really helpful in our attempts to explain the full meaning of Wittgenstein’s term. But it is true, and ought to be borne in mind, that Schiller’s insistence on freedom and the liberating features of Spiel chimes well with the fact that a fair number of Wittgenstein’s examples concern activities that are not obviously, or clearly not at all, rule-governed. And while such activities are evidently covered by the German word ‏”‎Spiel”, one would find it hard to make the English word ‏”‎game” do the same work.

My second point is just a reminder: Wittgenstein himself states that our ‏”‎language-games stand there as objects of comparison which, through similarities and dissimilarities, are meant to throw light on features of our language” (PI, 130). Now, if these objects of comparison are meant to illuminate factual uses of a language like German or English, they themselves must in some sense be non-consubstantial with these particular uses of language. Otherwise, they would not be fit to serve as objects of comparison, since comparison presupposes a certain degree of distinctness: one thing can only throw light on another if there is some distance between them. In other words, for such an object of comparison to do any useful work it will, for example, have to be strange or artificial or in some other way salient to let the contrast between it and what it is meant to illuminate do its clarifying work.

But what could be stranger or in a certain sense more artificial than the shopkeeper scene described in the opening remark of the Investigations? In particular, there are two activities sketched in a way that magnifies or augments them out of all proportion. The first of these activities is the shopkeeper’s way of using a colour chart and sample to identify the right sort of apples. The second one is his use of the series of number-words in counting up to the right number of apples.

Both activities are characterised in an absolutely exaggerated and unrealistic fashion. If one wanted to classify the scene, not in philosophical terms, but in terms of literary genre, one might wish to say that this is a piece of absurd theatre, perhaps a sketch written by Ionesco or Beckett.^[10] On the other hand, there clearly is a didactic element as well, so one might think of Brecht as a possible author. And if one expects the strikingness of the Verfremdungseffekt to be correlative with the explanatory power of this exercise, then this last ascription doesn’t seem all that unlikely.

At any rate, one could say that the strangeness and artificiality of the description are means to an instructive end. And to the extent they help us to identify, to memorise, and thus to repeat all the individual steps involved in the activities concerned, they are also at the service of what one might wish to call a ‏”‎formal” characterisation. So, if you look at the matter from a certain perspective, a description of the shopkeeper scene could be regarded as providing us with an in this sense ‏”‎formal” means of clarifying certain ways of using our natural language.

But what is really happening according to the description of the shopkeeper scene? At first one might be tempted to reply by saying that the grocer’s activities are decomposed into their elements and portrayed as a sequence of visible actions that in normal circumstances are in some sense performed but not perceived or not perceivable. This, however, is clearly not the case: an ordinary shopkeeper’s actions simply do not involve manipulations of colour charts and samples, nor will his counting apples involve any recitation of number words up to the right one. So Wittgenstein’s statement that ‏”‎It is in this and similar ways that one operates with words” must be ironical, unless his word ‏”‎operate” is taken, not in its usual and innocuous sense of ‏”‎employing” linguistic expressions, but in the extremely ambitious sense of ‏”‎spelling out the processes and techniques on which our use of the relevant words is based”. And as a matter of fact, in a context like the present one the German word ‏”‎operieren” could be used to convey both irony and at least rudiments of the ambitious sense.

But of course, not a shred of this ambitious sense is really expressed by a statement like ‏”‎The grocer selected five red apples and put them in a bag”. What Wittgenstein is really describing is a relatively complicated kind of stage performance representing ways of acquiring techniques of identifying and sorting objects that play a role in everyday processes of buying and selling. So what lends this peculiar air of strangeness and absurdity to Wittgenstein’s sketch is the apparent attempt to depict both the actual shopping process and the way some of the techniques involved in shopping are learnt, and to depict both sets of phenomena as if they formed part of one and the same sequence of actions.

The strangeness of Wittgenstein’s description is analogous to that of old paintings purporting to depict fragments of two or even more worlds at the same time. By this I don’t mean pictures like those showing the annunciation, for example – that is, the moment at which the angel tells Mary that she is to be the mother of Christ. In this case we are dealing with one world only, and the only reason for marvelling at this scene lies in the unusual presence of a member of a higher world. No, what I mean are pictures attempting to show more than one world at the same time, for example, a priest preaching to an absolutely mundane congregation of parishioners as well as a group of angels assembled in a corner of heaven, playing musical instruments. There is no interaction between these two scenes – that is, between the preaching and the performing angels –, and it is not even clear that we can sensibly speak of simultaneous goings-on. At any rate, the strangeness has something to do with the inexplicable co-presence of things that do not belong to the same logical space – that do not have a common frame of reference.

In a similar way we may find it difficult or feel unable to see the elements of the scene ostensibly described by Wittgenstein as belonging to the frame of reference of everyday shopping. Even if it is admitted that mastery of certain techniques is a presupposition of our coming to grasp and employ concepts and the words expressing them, this does not mean that the process of learning those techniques is in any way present when they are applied. On the contrary, it can be very important to keep processes of acquisition and processes of application strictly separate. At the same time, we may wonder whether it might not be one of the points of Wittgenstein’s obviously strange description of the shopping scene to bring out that the as it were logical multiplicity characteristic of a process of learning will tend to be quite different from that of a process of application.

One way of finding out about this sort of multiplicity would consist in identifying the sorts and number of constitutive elements involved in either kind of process. Looked at from this perspective, it will soon become clear that most ordinary uses of colour-words like ‏”‎red” don’t rely on the manipulation of colour charts and samples. On the contrary, appealing to samples, for instance, would in many cases slow down the proceedings and confuse the participants. On the other hand, when it comes to teaching colour-words or discussing difficult cases charts and samples can prove very helpful. If, for instance, I hesitate over the question whether to call the cover of a book orange or red, I may resort to a colour chart and base my decision on a comparison between relevant samples and my copy of the book. But here it must be remembered that situations of teaching the use of words or of discussing difficult cases are different in kind from what goes on in everyday situations of more or less successful ordinary uses of one of our natural languages.

Now what about the second activity described in the shopkeeper scene? That is, what about the way the grocer’s technique of using the series of number-words is described? Again we observe a kind of concertina effect, as the elements constituting an ordinary process of counting are pulled wide apart. One point of proceeding in this fashion may become clearer if we have a look at a distinction mentioned by Paul Benacerraf, even though to my mind his attempt at making the rationale of this attempted distinction plausible fails. He writes as follows:

There are two kinds of counting, corresponding to transitive and intransitive uses of the verb ‏”‎to count”. In one, ‏”‎counting” admits of a direct object, as in ‏”‎counting the marbles”; in the other it does not. The case I have in mind is that of the preoperative patient being prepared for the operating room. The ether mask is placed over his face and he is told to count, as far as he can. He has not been instructed to count anything at all. He has merely been told to count. (Benacerraf 1965/1983, 274)

I think the reason why Benacerraf’s attempt should be regarded as a failure is pretty obvious: in the case of intransitive vs. transitive verbs we are dealing with two classes of verbs that can usefully be distinguished according to whether they can or cannot admit of a direct object. In the case of the patient, however, we are not witnessing a second type of counting besides ordinary counting of the kind exemplified in counting coins in my pocket or pencils on my desk. The man ‏”‎counting” (in inverted commas) before being moved to the operating theatre does not count at all. He is merely reciting numerals in their canonical order. Even if he ‏”‎counted sheep”, as we say, he would not be counting in the true sense of the word for the simple reason that making and detecting mistakes are logically excluded, whereas the possibility of making and detecting mistakes is a condition of counting proper. Another point worth mentioning is this: it is obvious that uses of the word ‏”‎counting” in inverted commas are parasitic on there being cases of ordinary counting.

Now, what I have here tried to point out very briefly can nicely be mapped onto the shopping scene: the pronouncing of elementary number-words corresponds to the patient’s reciting numerals in their canonical order, but it is clear that this describes only part of what the grocer is doing; what is missing is the accompanying selection of apples from the drawer, which turns the whole process into one of counting apples. Of course, here the reciting of number-words stands in for any number of related processes like tacit counting, scanning objects, ticking their names on a written list, and so on. But at any rate, we are only dealing with a full-fledged case of counting if the recitation is supplemented by a process of correlating objects, or sets of objects, with those number-words.

However, as Wittgenstein himself points out, this does not stand in the way of conceding an independent role to the activity of reciting number-words. Thus, in learning their mother tongue children will play all kinds of preparatory games with words that will eventually lead to mastery of their full use. Reciting number-words in their canonical order is one of these games, but it isn’t, as Wittgenstein emphasises, a full-fledged use of linguistic expressions. It is, as he calls it (in PI, 7), ein sprachähnlicher Vorgang – a ‏”‎speech-like” or ‏”‎quasi-linguistic” activity. The goal of learning one’s first language gives these ‏”‎primitive” language-games their point, but it is not sufficient to turn them into fully developed uses of words.

Before getting down to sketching an answer to my earlier question about the status of the shopkeeper scene, I should like to touch on two points directly relevant to some of the things we have said about Wittgenstein’s notion of a language-game. The first of these points is connected with his description of the act of shopping in terms of charts, samples, and written signs. Of these samples, etc. he goes on to say that ‏”‎It is most natural, and causes least confusion, if we count the samples as tools of the language” (PI, 16). Here it should be remembered that the meaning, or meanings, of the German word ‏”‎Muster” are not fully captured by the English word ‏”‎sample”. For in many cases a Muster is primarily a model in the sense of a paradigm, an exemplar, a paragon. It is something that stands out as calling for imitation.

It is obvious that most of these samples or paradigms do not form part of language in a narrow sense: excepting a small number of special cases,^[11] they don’t stand on the same level, and are not of the same sort, as words. But they are all the same very closely bound up with language, and the tie that connects them is a practical one: in learning to use colour-words, for example, we also learn to manipulate colour charts or, at any rate, we learn to apply certain techniques that would allow us to handle such charts and the samples supplied by them. And, on the other hand, in learning to recite number-words we take a first step towards learning to count.

Clearly, one of the points of the strangely magnified and augmented scene described in Wittgenstein’s description of the shopkeeper scene consists in bringing out the fact that a fairly complicated practice involving a great variety of abilities needs to be understood and mastered if one is to be able to perform what seem to be extremely simple actions like counting or selecting a certain type of apples.

Only if one has fully assimilated the idea of our having to acquire those (or a different cluster of) background techniques and to possess a great variety of largely implicit abilities will one be able to make much of the following passage:

We can, however, also imagine such a [colour] chart’s being a tool in the use of the language. Describing a complex is then done like this: the person who describes the complex has a chart with him and looks up each element of the complex in it and passes from the element to the sign (and the person to whom the description is given may also translate its words into a picture of coloured squares by the the use of a chart). This chart might be said to take over here the role that memory and association play in other cases. (PI 53)

Now, if the the chart can in such cases be claimed to play the role of memory and association, then it can to this extent be ascribed the capacity to replace memory and association in our descriptions of relevant situations. And if this is a legitimate move, then it will be seen that this can have an important bearing on arguments in the philosophy of mind, for instance. Once we see this, we also understand that this is a move really made by Wittgenstein in the context of the private-language argument and related parts of his work. What I want to underline at this point is the fact that this manoeuvre grows out of considerations Wittgenstein develops in the context of describing and examining primitive language-games and their tools. It represents one way in which the invention, description, and examination of language-games can prove philosophically useful.

The second point I wish to make here will be put even more briefly than my first point, and to some extent it will be parallel to the first one. At a crucial stage of the private-language argument developed in the Investigations Wittgenstein uses an idea which heavily relies on a specific and powerful notion of a language-game. This is the idea that such games have beginning and end, which in my view, however, doesn’t mean that in every relevant case one particular starting- and end-point has to be spelled out in a clear-cut way. What it does mean is that the type of situation we are interested in will have to be compatible with a description in terms of a game with beginning and end. To illustrate this idea I shall quote part of the well-known remark 290 from the Investigations:

[1]

It is not, of course, that I identify my sensation by means of criteria; it is, rather, that I use the same expression. But it is not as if the language-game ends with this; it begins with it.

[2]

But doesn’t it begin with the sensation – which I describe? – […][12]

The second position characterised by this schematic exchange is that of a philosopher who claims that his utterance of ‏”‎I am in pain” is as it were the second move in a game that begins with a presentation of the sensation of pain itself. But once he conceives of the game in this fashion, the second move will naturally be read as involving a description of what was presented by the first move. If, on the other hand, you see the game as beginning with an utterance of ‏”‎I am in pain”, it will be natural to read these words as a kind of avowal – as a natural expression of pain and not as a description of a given sensation.

The second game – the game allegedly beginning with the sensation – might be believed to terminate in a characterisation of the relevant sensation as pain (as opposed to hunger, for example, or nostalgia or surprise). Such an attempt at a characterisation, however, would involve the idea that a sensation of pain, for instance, could be identified independently of its being felt as pain – and, according to Wittgenstein, this idea might lead us to hold all kinds of absurdity.

The first game, however – that is, the game beginning with the pain-expression –, would have to continue beyond this point: a probable second move might for instance consist in another person’s response to my utterance ‏”‎I am in pain”. One typical response would amount to inquiring into the source of the first speaker’s complaint, trying to look after him and to alleviate his suffering. Another response might consist in the short reply: ‏”‎Come on, you look absolutely fine. I’m sure that there is nothing really wrong with you.” And so on.

So the whole structure of the description of a given situation and of the ensuing philosophical discussion will depend on which of these two language-games seems more suitable. If you opt for the game beginning with the sensation itself, you will be obliged to make sense of the idea of reading the words ‏”‎I am in pain” as a description. If you opt for the game beginning with the pain-expression, you will be obliged to make it plausible to your interlocutor that ignoring the sensation itself can be philosophically advantageous.

This is the point where my paper is drawing to a close. Let me just try to give a tentative answer to the question why Wittgenstein may have decided to change the numbering of the first two language-games – the shopkeeper scene and the language of the builders. I think that one fairly obvious and important difference between the two lies in the fact that the builders’ game (as opposed to the shopping scene) cries out for extension. Its primitiveness is not final; it is, among other things, meant to invite readers to think about fruitful ways of constructing a more complicated and more powerful kind of machinery. And as we know, this is exactly the sort of question Wittgenstein discusses in later parts of the book, and occasionally with explicit reference to the builders. The shopping scene, on the other hand, does not cry out for extension or completion at all. It is, in a sense, over-complete inasmuch as it displays more structure than we should usually want to perceive. So this is one reason for not treating the two language-games as games having equal status.

The second difference I wish to point out involves a certain amount of speculation. Nominally, the account of the builders’ game describes an application of Augustine’s story. The shopping scene, however, doesn’t mention him and doesn’t seem to make any reference to the introductory quotation at all. But in my view it could profitably be seen as a reply to the quoted passage. In a way, Augustine’s story is too short and leaves all the questions raised by it open. It as it were poses as an explanation but only succeeds in suggesting new questions concerning, for example, the unity of our concept of meaning and the functioning of ostensive explanations.

The shopkeeper scene, on the other hand, is a reply to Augustine’s story – only that I am not sure whether to call it a direct or an indirect one. In contrast to Augustine’s account, which is too short, the story we are told is too long. It does show that operating with words involves mastery of various techniques, but by being so over-explicit and tediously precise it at the same time helps to get two messages across. The first of these is enshrined in the suggestive and studiously non-imperatival statement that explanations come to an end. And here it is important to remember that Wittgenstein doesn’t say that explanations have to come to an end: they come to an end – full stop. The second is the claim that, instead of talking about meanings, we should often content ourselves with talking about the use of words. Both messages may gain in plausibility by being seen in direct contrast with Augustine’s story. So this may be a second reason why Wittgenstein decided to add the shopping scene to the story told by St Augustine.